Technologies and Innovation – Decision Making

Technologies and Innovation – Decision Making

• 1.
  Digital Technologies andInnovation Decision-Making April 2019 http://DSign4Methods.com
• 2.
  2 Innovation is aState of Mind ©2019 LHST sarl Introduction Session 1 The Building Blocks Session 2 Innovation Session 3 Digital Economics Session 4 The Internet of Value Session 5 Decision Making Session 6 Data Ethics
• 3.
  ©2019 L. SCHLENKER Agenda Introduction TheData Revolution Time, Space and Organization The Analytical Method Introduction
• 4.
  Productivity • Harder, better,faster… • Mechanized productivity • Knowledge productivity • Continuous Productivity Steven Sinofsky ©2019 L. SCHLENKER Introduction
• 5.
  • Reinforced learningand predictive analytics • “ Algorithms tailor search to our interests far better than any one human ever could” • Merge Sort, Fourrier Transform, RSA algorithm, PageRank (eigenvalues)… • The UK’s the Alan Turing Institute “Informally, an algorithm is any well-defined computational procedure that takes some value, or set of values, as input and produces some value, or set of values, as output.” Introduction to Algorithms 3rd edition. ©2019 L. SCHLENKER Introduction
• 6.
  – Intelligence: Theidentification of a challenge that requires data collection and a relevant decision – Design: Exploring, planning, and analyzing alternative courses of action – Choice: Selecting the appropriate course of action H. A. Simon Introduction
• 7.
  Copyright © 2004by South-Western, a division of Thomson Learning. All rights reserved. Introduction
• 8.
  Information Technology Business Analytics Decision Making INDIVIDUALS GROUPS ORGANIZATIONS • Programmed Decisions –Situations in which past experience permits decision rules to be developed and applied in the future • Non_programmed Decisions – responses to unique, poorly defined challenges that have important consequences to the organization Introduction
• 9.
  UNCERTAINTY • Facts notknown • Look for Information • Fact Finding /.Analysis DATA BASED COMPLEXITY • Too many facts • Produce Information • Simulation/Synt hesis MODEL BASED EQUIVOCALITY • Facts not Clear • Analyse Information • Application of Expertise KNOWLEDGE BASED Introduction
• 10.
  How do greatleaders make great decisions? Introduction
• 11.
  System 1, System2 • Thinking Fast, Thinking Slow (2011) presents a dichotomy between two modes of thought • "System 1" is fast, instinctive and emotional (intuition) • "System 2" is slower, more deliberative, and more logical (reasoning). • Biases have two sources of error, the observed behavior and “rationality” Introduction
• 12.
  • In the'simple' domain, problems and solutions are known. There is a one-to- one relationship between cause and effect. • In the 'complicated' domain, problems and solutions are knowable. There is a one to N relationship between cause and effect. • In the 'complex' domain, problems or solutions are unknown. There is a N to N relationship between causes and effects. Snowden and Boone Introduction
• 13.
  – Intelligence: Theidentification of a challenge that requires data collection and a relevant decision – Design: Exploring, planning, and analyzing alternative courses of action – Choice: Selecting the appropriate course of action H. A. Simon Jack is looking at Anne, but Anne is looking at George. Jack will be successful, but George will not. Is a successful person looking at an unsuccessful one? ©2019 L. SCHLENKER Introduction
• 14.
  • Cost -It is often less costly to analyze decision problems using models. • Time - Models often deliver needed information more quickly than their real-world counterparts. • Comprehension - Models can be used to do things that would be impossible. • Models give us insight & understanding that improves decision making. Introduction
• 15.
  Why do wetake poor decisions? • The object of measurement (i.e., the thing being measured) is not understood. • The concept or the meaning of measurement is not understood. • The methods of measurement are not well understood ©2019 L. SCHLENKER Introduction
• 16.
  • What doesproductivity mean (faster, more impressive, more precise) ? • Is it observable – how is something more precise answer to a problem? • The challenge is deciding what we want to measure Lewis Mumford, Technics and Civilization • What does productivity mean (faster, more impressive, more precise) ? • Is it observable – how is something more precise answer to a problem? • The challenge is deciding what we want to measure Introduction
• 17.
  • Reducing thenumber of potential outcomes is the key to better decision-making • Develop unambiguous definitions and measurement • What data do I have, Choose the appropriate measure • Understand how people react to the data www.google.com/dashboard ©2019 L. SCHLENKER Ask Examples Resources Is it possible that this may already have been researched? The average cost of IT training for given type of user Go to the library (Internet) Could it be projected from past experience? Growth in product demand Research the market Does it leave a trail of some kind? Current level of customer retention Look for the data Could it be observed in real-time? The amount of time an equipment operator spends filling out forms Unsupervised learning Can it be tested? The effect of a new system on the productivity of a sales clerk Supervised learning Introduction
• 18.
  Introduction
• 19.
  Types of Learning •Supervised (inductive) learning Training data includes desired outputs • Unsupervised learning Training data does not include desired outputs • Semi-supervised learning Training data includes a few desired outputs • Reinforcement learning Rewards from sequence of actions Introduction
• 20.
  Supervised (inductive) learning •Training data includes desired outputs Unsupervised learning • Training data does not include desired outputs Semi-supervised learning • Training data includes a few desired outputs Reinforcement learning • Rewards from sequence of actions ©2018 L. SCHLENKER Introduction
• 21.
  Introduction
• 22.
  • Decision-tree modelsoffer a visual tool that can represent the key elements in a model for decision making • Decision trees are a comprehensive tool for modeling all possible decision options. • While influence diagrams produce a compact summary of a problem, decision trees can show the problem in greater detail.  Supervised  Categorical It’s sunny, hot, normaly humid, and windy – should I play tennis? Introduction
• 23.
  • Cluster analysisis used for identifying natural groupings of things. • Data instances that are similar to each other are categorized into one cluster. • Similarly, data instances that are very different from each other are moved into other clusters. • Clustering is an unsupervised learning technique as there is no output or dependent variable  Unsupervised  Discrete ©2018 L. SCHLENKER Introduction
• 24.
  • Associate rulemining, or market basket analysis; is a popular, unsupervised learning technique, used in business to help identify shopping patterns. • It helps find interesting relationships (affinities) between variables (items or events). • Thus, it can help cross-sell related items and increase the size of a sale. • There is no dependent variable – and no right of wrong answer “A Customer who bought bread an butter also bought a carton of milk 60 percent of the time.“  Unsupervised  Categorical Introduction